All news

Waymo's first paid rides make robotaxis a contest of fleet capital

A long row of smooth graphite pod shapes in a pale concrete hall, the nearest one in cobalt

Waymo began charging riders in Phoenix with a safety driver in every car. Self-driving now competes on who can fund fleets, not only on software.

Waymo began charging for self-driving rides in the Phoenix suburbs on 5 December, opening its Waymo One app to a few hundred riders from its research program across Chandler, Tempe, Mesa and Gilbert. A Waymo-trained driver still sits behind the wheel of every car.

The fare matters more than the scale. Once riders pay, self-driving stops being a research line and becomes a transport business, and transport businesses are won by whoever can finance the most vehicles at the lowest cost per mile. Driving software that works is now the entry ticket. What follows is a contest of fleet capital: cars, sensors, depots, insurance and the staff who support riders remotely, all paid for years before a network earns back its cost.

Waymo has lined that capital up. This year it signed agreements for up to 62,000 hybrid minivans and up to 20,000 electric SUVs, and it has said the electric fleet alone could handle up to a million trips a day. Its parent carries the bill. Alphabet's Other Bets segment, which holds Waymo alongside the group's other long-term ventures, lost $727 million in the third quarter on revenue of $146 million. The closest rival is financed the same way: Cruise, the self-driving unit of General Motors, reached a $14.6 billion valuation in October after a Japanese carmaker committed $2.75 billion, on top of $2.25 billion from a financial investor in May.

Those sums put a full robotaxi stack out of reach for a start-up. A company that must raise a billion dollars to buy cars before its first fare is a carmaker's subsidiary in waiting. The openings sit in what every fleet operator will have to buy: simulation that cuts the miles a car must drive in testing, cheaper sensors, maps kept current, software that lets one remote operator watch many cars, and the claims data insurers need to price a mile with no driver. Fanaura judges that layer by the test at the core of our thesis, whether the customer's cost per unit falls as the product scales, and founders building it can expect that question first.

Skeptics point at what has not changed. The drivers are still in the seats, the riders are pre-screened volunteers, and the Phoenix suburbs offer wide roads, little rain and few pedestrians. Until the safety driver leaves, every paid trip carries the cost of a human and the hardware of a self-driving car at once, and a fleet of 82,000 vehicles would multiply losses rather than profits. A date for paid rides with an empty driver's seat, and an opening of Waymo One beyond its early riders, would settle the argument. Until both arrive, the capital race runs ahead of the economics it is meant to secure.

Self-driving has a price list now, and the first line on it is the cost of capital.

For founders

Building simulation, remote operations or insurance data for autonomous fleets? Send us the project.

Fanaura invests its own capital and the capital of its limited partners (LPs), usually through a separate SPV for each deal. If we see a fit, we get in touch.

Apply in five minutes
Subscribe

Get the news by email

One short note each weekday. No marketing, and you can unsubscribe at any time.